An example of library and museum cooperation through the FRBROO conceptual model
•Télécharger en tant que PPT, PDF•
3 j'aime•1,998 vues
Paper given on the occasion of the CIDOC Annual Conference, 'Enriching Cultural Heritage,' Helsinki, 10-14 June 2012
Recommandé
Recommandé
Contenu connexe
Tendances
Tendances (20)
Similaire à An example of library and museum cooperation through the FRBROO conceptual model
Similaire à An example of library and museum cooperation through the FRBROO conceptual model (20)
Dernier
Dernier (20)
An example of library and museum cooperation through the FRBROO conceptual model
- 1. An example of library and museum cooperation: the FRBROO conceptual model Patrick Le Boeuf National Library of France CIDOC Annual Conference – 'Enriching Cultural Heritage' – Helsinki, 10-14 June 2012
- 2. Museums and libraries akin but different… mostly unique objects mostly copies of publications How could they cooperate (at least as far as cataloguing is concerned)?
- 3. Museums and libraries different but akin… also non-unique objects also unique objects … and interrelated objects
- 4. Museums and libraries It does make sense to strive to make museums' and libraries' databases interoperable through a common conceptual model
- 6. A striking coincidence • 1998: release of a preliminary version of CIDOC CRM by ICOM CIDOC • 1997: FRBR model approved by IFLA
- 7. FRBR and the 'FR family' • FRBR = Functional Requirements for Bibliographic Records (1991-7, published 1998) – data pertaining to publications and holdings • FRAD = Functional Requirements for Authority Data (published 2009) – data pertaining to authors and works • FRSAD = Functional Requirements for Subject Authority Data (published 2011) – data pertaining to subjects and classification numbers
- 8. CIDOC CRM • CRM = Conceptual Reference Model • 1996: beginning of work • 2000: formation of CIDOC CRM SIG (Special Interest Group) • 2006: ISO standard ISO 21127:2006
- 9. FRBR/CIDOC CRM Harmonisation Group • Formed 2003 • Representatives from – IFLA FRBR Review Group – CIDOC CRM SIG • Objective: develop a unified model for descriptions of both unique and non- unique objects
- 10. How? • Early decision: CIDOC CRM as reference point • 'Translate' FRBR into the CIDOC CRM formalism • Examine each attribute and relationship defined in FRBR • Find equivalents in CIDOC CRM • Where no equivalent, declare new subclasses and subproperties
- 11. How? • Refine FRBR notions deemed too vague • Add some new classes and properties to CIDOC CRM • 2003-2012: 19 meetings • June 2009: version 1.0 of FRBROO is released (OO = 'object-oriented') • 2009-2012: work on FRAD and FRSAD • 2012: version 2.0 is due
- 12. The FR family An introduction
- 13. FRBR, or, How to make a catalogue • FRBR entities reflect the way cataloguers organise their work • First, they examine a physical object…
- 14. FRBR, or, How to make a catalogue • Then, they extrapolate the characteristics of the publication the physical object belongs to Publication Publication Publication Publication
- 15. FRBR, or, How to make a catalogue • To end with, they use uniform titles to collocate publications with interrelated contents (e.g., the publication of a text and the publication of a translation of that text) Publication Publication Content Content Family of contents
- 16. FRBR, or, How to make a catalogue • Work • Expression • Manifestation • Item Publication Publication Content Content Family of contents
- 17. The Manifestation problem • Manifestation = abstract definition (intension) of a set of items • When a set has only one element, how to distinguish between intension and extension?
- 18. The Manifestation problem No 'Manifestation' class in FRBROO • Instead, two disjoint classes: E28 Conceptual Object E55 Type F3 Manifestation Product Type E72 Legal Object E24 Physical Man-Made Thing F4 Manifestation Singleton
- 19. Work and Expression • A distinction that did not exist in CIDOC CRM • E73 Information Object = both concepts and their expression CIDOC CRM had to be modified • May 2008 (v. 4.2.5): introduction of E89 Propositional Object and E90 Symbolic Object
- 20. Work and Expression E89 Propositional Object F1 Work E73 Information Object E90 Symbolic Object F2 Expression
- 21. Other FRBR entities Expression Manifestation Item Person Corporate Body Concept Object Event Place E39 Actor Work 'is about' (= E89 Propositional Object. P129 is about (is subject of) E1 CRM Entity
- 22. Authority work • Consists of – managing various appellations for any kind of thing – stating relationships among those things • Is modelled in FRAD and FRSAD • Will be addressed in FRBROO v. 2.0 Michelangelo Buonarroti Michael Angelo Michelangiolo Michel-Ange Michał Anioł Buonarroti, Michelagniolo
- 23. Authority work and CIDOC CRM • Some aspects of authority work were deemed important enough to be introduced in CIDOC CRM • Version 4.2.2 of CIDOC CRM (August 2007): E42 Object Identifier was renamed E42 Identifier, so as to apply to any kind of thing, like in the 'FR family'
- 24. Various types of Works How the Work notion was refined in FRBROO
- 25. The Work notion in FRBR Work = 'a distinct intellectual or artistic creation'
- 26. The Work notion in FRBROO F1 Work a sum of concepts… F14 Individual Work F15 Complex Work … uniquely and completely expressed in a single 'text' … expressed in a constellation of interrelated works
- 27. The boundaries of a 'complex work' • Always somewhat arbitrary… – are these two distinct works, or just one work? – With FRBR, such a question could lead to endless discussions… – With FRBROO and F15 Complex Work it gets easier
- 28. Complex works in museums as well? F14 Individual Work F15 Complex Work F14 Individual Work R10 has member (is member of) R10 has member (is member of) Piranesi, Carceri d'invenzione, plate XIII 1st state 2nd state
- 29. A third subclass of F1 Work F1 Work F21 Recording Work
- 30. A fourth subclass of F1 Work F1 Work F16 Container Work F17 Aggregation Work F19 Publication Work F20 Performance Work = work that adds value to expressions of other works by… … putting them together (e.g., an anthology) … publishing them … or performing them
- 31. Authority control A new feature of FRBROO v. 2.0
- 32. What is authority control? • An activity that consists of establishing controlled access points to uniquely identify any thing deemed to be of interest for users, and ensure it can be retrieved under its other forms of appellation Leonardo, da Vinci, 1452-1519 Леонардо, да Винчи, 1452-1519 Dürer, Albrecht, 1471-1528. Kleine Passion Dürer, Albrecht, 1471-1528. Gospel for the unlearned Mural painting and decoration Wall-painting
- 33. A model for authority control E41 Appellation F50 Controlled Access Point F13 Identifier F12 Nomen F35 Nomen Use Statement F52 Name Use Activity F34 KOS R37 states as nomen R 35 is specified by R32 is warranted by R64 used name 4000_ |a Леонардо, |c да Винчи, |d 1452-1519 Леонардо, да Винчи, 1452-1519 Леонардо да Винчи Library of Congress Authorities
- 34. Authority data and interoperability Buonarroti, Michelangelo Michelangelo Buonarroti, 1475-1564 Common URI (see also ICOM recommendation on Linked Open Data for museums, http://network.icom.museum/fileadmin/user_upload/minisites/cidoc/AGM_2011/LoD_Fo )
- 35. Introducing events in FRBR How works, expressions, manifestations, and items come into being
- 36. The CIDOC CRM Top-level classes useful for integration participate in E39 Actors E55 Types E28 Conceptual Objects E18 Physical Thing E2 Temporal Entities E41Appellations affect or / refer to refer to / refine referto/identify location atwithin E53 Places E52 Time-Spans
- 37. E65 Creation F27 Work Conception R16 initiated (was initiated by) F1 Work F2 Expression F28 Expression Creation E12 Production R3 is realised in (realises) R17 created (was created by) F_ Manifestation Creation?
- 38. F3 Manifestation Product Type F4 Manifestation Singleton E83 Type Creation CLR6 should carry (should be carried by) F30 Publication Event F24 Publication Expression E12 Production F28 Expression Creation P135 created type(was created by) R24 created(was created through) P108 produced (was produced by) R18 created (was created by)
- 39. F32 Carrier Production Event R28 produced (was produced by) E12 Production F5 Item F3 Manifestation Product Type R7 is example of (has example)
- 40. Introducing events in FRBR Other types of events
- 41. F21 Recording Work F29 Recording Event F26 Recording F28 Expression Creation F33 Reproduction Event E84 Information Carrier E84 Information Carrier E55 Type {Reproduction} R22 created a realisation of (was realised through) E12 Production R21 created (was created through) R13 is realised in (realises) R29 reproduced (was reproduced by)R30 produced (was produced by) P130 shows features of (features are also found on) P130.1 kind of similarity
- 43. Introducing events in FRBR Events relating to authority control
- 44. F40 Identifier Assignment E42 Identifier R46 assigned (was assigned by) F13 Identifier F50 Controlled Access Point E1 CRM Entity E90 Symbolic Object E15 Identifier Assignment R45 assigned to (was assigned by) P142 used constituent (was used in) Леонардо, да Винчи, 1452-1519 Леонардо
- 45. E41 Appellation F35 Nomen Use Statement F52 Name Use Activity R32 is warranted by (warrants) R64 used name (was name used by) F24 Publication Expression P67 refers to (is referred to by) Леонардо да Винчи
- 46. F42 Representative Expression Assignment R50 assigned to (was assigned by) F41 Representative Manifestation Assignment F15 Complex Work F2 Expression F3 Manifestation Product Type F4 Manifestation Singleton R51 assigned (was assigned by) R48 assigned to (was assigned by) R49 assigned (was assigned by) R53 assigned (was assigned by)
- 47. What remains to be done
- 48. What after FRBROO v. 2.0? • Two types of resources were only partially modelled in FRBR: – Continuing resources – Digital resources Partially modelled in FRBROO as well
- 49. What after FRBROO v. 2.0? • Digital resources – See if CRMdig could be used in combination with FRBROO? • Continuing resources = – Serials (periodicals, bibliographic series) • Some classes and properties to be added – Integrating sources (loose-leaf publications, updating digital resources) • To be checked
- 50. The work notion in libraries and museums
- 51. 'Work' in libraries and museums • Librarians would presumably feel urged to 'collocate' these 'expressions' through a reference to an alleged 'common work:' • In museums this is not the case • F15 Complex Work supports both views
- 52. To conclude From CIDOC CRM and FRBROO to Linked Data
- 53. Library and museum cooperation in the time of Linked Data • No need to 'merge' our databases • FRBROO and CIDOC CRM can help integrate the knowledge contained in our databases… • … provided we agree on common rules for URI assignment! • Linked Data is an opportunity for increased cooperation between libraries and museums
Notes de l'éditeur
- Museums and libraries are at the same time close relatives and rather different entities. Both types of institutions share the same mission of preserving cultural heritage and the same responsibility when it comes to producing, storing, and disseminating information about that cultural heritage. On the other hand, the collections held in one and the other display characteristics that make it, at first sight, difficult to envision a common framework for their description. Most library holdings are non-unique copies of ‘publications,’ in the broadest sense of the term, i.e., non-unique exemplars of products obtained as the result of more or less industrialised processes, while museums are mainly concerned with unique items. Therefore, although museum and library cooperation has been increasingly presented as highly desirable for a number of years, there are few practical examples of such cooperation, at least in the field of cataloguing.
- However, the boundaries between the respective typologies of museum and library collections are less sharp than it seems. Libraries do also hold unique items, such as manuscripts, and museums may also have to deal with non-unique objects, such as natural history specimens, and art and photograph prints. Besides, libraries hold resources that are about museum items, and museums hold resources that are related in some way to library items.
- It does therefore make sense indeed to strive to cooperate and find a way to make our databases interoperable through a common conceptual model.
- It is a striking coincidence that the library community and the museum community endowed themselves almost simultaneously with conceptual models for the information they produce. IFLA, the International Federation of Library Associations and Institutions, approved the FRBR model in 1997 and had it published in the following year. ICOM CIDOC released a preliminary version of CIDOC CRM in 1998.
- The acronym FRBR stands for: Functional Requirements for Bibliographic Records , which was the title of a study conducted from 1991 to 1997 by an IFLA Group in order to answer the following question: ‘Can cataloguing be considerably simplified?’ FRBR only deals with the information conveyed by bibliographic records and holdings records; the information contained in authority records is addressed in a second model that was developed later, FRAD (for: Functional Requirements for Authority Data ), and subject relationships are the focus of a third model, FRSAD ( Functional Requirements for Subject Authority Data ), which is the youngest member of the so-called ‘FR family.’
- As all CIDOC members will know, work on the CIDOC CRM began in 1996. In 2000, the CIDOC CRM Special Interest Group (CIDOC CRM SIG) was formed for the sole purpose of developing the model further. CIDOC CRM became in 2006 an ISO standard, ISO 21127.
- It was not until 2003 that a joint group was officially established to the purpose of harmonising museums’ and libraries’ conceptualisations. The International FRBR/CIDOC CRM Harmonisation Group was formed as both a working group affiliated to the IFLA FRBR Review Group, and a subgroup of the CIDOC CRM SIG. One of the first tasks of the Harmonisation Group was to list the differences between library information and museum information, and define the challenge of developing a unified model for descriptions of both unique and non-unique items.
- From the very beginning, it was decided that CIDOC CRM would be used as our reference point, and that its formalism would be the ‘target language’ into which FRBR was to be ‘translated.’ Basing on that principle, the Group examined all attributes and relationships declared in the original definition of FRBR, and strove to extract their semantics. Then the Group’s members expressed each attribute and relationship in the CIDOC CRM formalism – i.e., re-using a CIDOC CRM structure every time it seemed feasible, or declaring new subclasses and subproperties of extant CIDOC CRM classes and properties in the contrary case. Each of the ‘attributes’ declared in FRBR was thus transformed into one or more than one ‘property’ in the resulting model, which was named FRBROO, where OO refers to the transformation from entity-relationship to object-oriented formalism.
- Both models were impacted by this process. Some notions in the FRBR conceptualisation were deemed too vague to be plugged to CIDOC CRM such as they stood, and had to be refined. On the other hand, some new classes had to be introduced in CIDOC CRM, so that they could be declared as super-classes for some concepts defined in the FR family. Eventually, some notions inherent to library practice were deemed sufficiently relevant for museums for them to be included in CIDOC CRM itself rather than its extension FRBROO. From 2003 to 2009, the Harmonisation Group worked on the ‘translation’ of the FRBR model, which led to the release of FRBROO version 1.0 in June 2009. From 2009 to 2012, the group focused on the integration of FRAD and FRSAD, i.e., notions relating to authority data. This work was completed during the group’s latest meeting. Version 2.0 of FRBROO, containing the conceptualisation of the entire FR family, should be released in the near future.
- FRBR reflects the way cataloguers organise their work. They first receive a physical thing (a book, a map, a score, a DVD, a CD…), and they examine it.
- They extrapolate from that examination a number of characteristics which should normally be found on any other copy of the same publication; as a consequence, the bibliographic record they create is reputed to describe the publication itself.
- In some cases, they will provide that bibliographic record with a uniform title that serves to collocate all the various publications that happen to contain the same intellectual production, and they will provide that uniform title with some additional elements (e.g., a statement of the language used) that serve to differentiate among those publications.
- In FRBR parlance, the physical thing is an instance of the Item entity, the publication is an instance of the Manifestation entity, the abstract thing referred to by the differentiating combination of the uniform title and its additional elements is an instance of the Expression entity, and the abstract thing referred to by the collocating uniform title is an instance of the Work entity.
- In the case of unique products such as manuscripts, the distinction between Manifestation and Item becomes blurred. The Manifestation entity actually matches the mathematical notion of ‘a set:’ an instance of Manifestation consists of the abstract definition of a set of items. As any mathematical set, a manifestation can have many elements, or just one. When it has only one element, it becomes difficult, in practice, to draw the line between the abstract definition of the characteristics of the set, and the unique physical carrier that displays all those characteristics.
- Consequently, FRBROO did not recognise the notion of Manifestation per se, but split it into two distinct (and disjoint) classes, F3 Manifestation Product Type (the abstract type exemplified by any copy of a given publication) and F4 Manifestation Singleton (a unique physical object, such as a manuscript).
- The distinction between Work and Expression was something totally new for the CIDOC CRM conceptualisation. When the harmonisation process began, both notions could only be mapped to E73 Information Object, which covers indifferently the ideational features of an intellectual product and the realisation of that intellectual product. CIDOC CRM had therefore to be modified. This is the reason why two new classes were introduced in the model: E89 Propositional Object and E90 Symbolic Object. An instance of E89 Propositional Object consists solely of concepts, and an instance of E90 Symbolic Object consists of signs.
- E73 Information Object is declared as a subclass of both E89 Propositional Object and E90 Symbolic Object. In FRBROO, F1 Work is declared as a subclass of E89 Propositional Object only, and F2 Expression as a subclass of E73 Information Object (and therefore, indirectly, as a subclass of both E89 Propositional Object and E90 Symbolic Object). This shows that the harmonisation process involved actual cooperation between libraries and museums: librarians accepted to correct the oversimplifications of their model and to transform it into an extension to CIDOC CRM, while representatives for the CIDOC accepted to make the CIDOC CRM evolve in order to accommodate the FRBR model.
- Any of the aforementioned four entities – Work, Expression, Manifestation, and Item – can be connected to either Person or Corporate Body, two entities that are subsumed in the CIDOC CRM notion of E39 Actor. Besides, the Work entity can be connected to any other entity through the ‘subject relationship,’ meaning that a given work can be about any instance of any entity. This relationship corresponds to the CIDOC CRM property P129 is about (is subject of) from E89 Propositional Object to E1 CRM Entity.
- The so-called ‘authority work,’ which forms a significant part of the cataloguing process, consists of managing the various appellations that serve to refer to instances of any of the aforementioned entities, and stating the relationships that exist between and among instances of some of those entities (e.g., between two instances of Person, or between a person and a corporate body, etc.).
- The process through which standardised forms of appellations are created and assigned to any kind of things was deemed useful enough in the museum context for it to be imported in CIDOC CRM. This is the reason why class E42, which was originally labelled Object Identifier, was renamed just Identifier in version 4.2.2 of CIDOC CRM (August 2007); originally, it covered mainly inventory numbers, and was extended to any kind of conventional way of referring uniquely to any kind of thing.
- The original version of the FRBR model recognised a single Work entity, which served to cover any kind of ‘distinct intellectual or artistic creation.’
- While FRBROO retains a very broad and generic class named F1 Work that corresponds to the original Work entity, the Harmonisation Group felt the need for some refinements that can prove useful for the description of museum items as well. FRBROO distinguishes between two notions that are declared as subclasses of F1 Work: F14 Individual Work, and F15 Complex Work. An individual work is the sum of concepts that is uniquely and completely expressed in a given ‘text’ (in the broadest sense of that term), while a complex work is a constellation of works that are deemed sufficiently related to each other, in a given cultural environment, to be reputed to form a single entity or ‘family.’
- Inevitably, determining the boundaries of any complex work is largely an arbitrary process: are a novel and the movie based on that novel members of the same complex work, or not? In the context of the original FRBR model, such a question used to lead (and occasionally still does) to lengthy, heated, and perfectly pointless discussions. In FRBROO this difficulty is solved, since any member of a complex work can itself be a complex work.
- The notion of complex work can be useful in museum practice as well, e.g. in the case of art prints that exist in several distinct states. Using the CIDOC CRM alone, such a case can only be modelled through property P130 shows features of (features are also found on), which is not entirely satisfying, or through the explicit instantiation of the activity of E11 Modification that resulted in the alteration of the plate. Using FRBROO, all states of an art print are modelled as instances of F14 Individual Work that are members of a single instance of F15 Complex Work that serves to identify the art print as an abstract documentation unit comprising all its various states.
- Also, FRBROO distinguishes a third subclass of F1 Work: F21 Recording Work. It covers works the essence of which consists of capturing some aspects of a fleeting moment, e.g. through photographing, filming, or making sound recordings. Once again, such notions can be useful in museums as well.
- The fourth subclass of F1 Work declared in FRBROO is F16 Container Work, a notion that covers works the essence of which consists of adding value to expressions of other works through various processes, e.g. by juxtaposing them, as in an anthology (this notion is specialised as F17 Aggregation Work), or by editing them and providing them with an appropriate layout, as publishers do with the texts they publish (this notion is specialised as F19 Publication Work), or by performing them (this notion is specialised as F20 Performance Work). In museum practice, F17 Aggregation Work can be a useful class, when it comes to collages or works of appropriation art.
- As mentioned earlier, version 2.0 of FRBROO, which will be released in the course of 2012, will be augmented with a section on authority data. Authority control consists of establishing controlled access points that serve to uniquely identify any thing that is deemed to be of interest for users, and to ensure that such things can be retrieved under other forms of appellation under which they may happen to be known. Controlled access points can be created for persons or corporate bodies (e.g., ‘Leonardo, da Vinci, 1452-1519’ which identifies a person who can also be searched under ‘Леонардо, да Винчи, 1452-1519’), for products of the mind (e.g., ‘Dürer, Albrecht, 1471-1528. Kleine Passion’ which can also be searched under ‘Dürer, Albrecht, 1471-1528. Gospel for the unlearned’), or for subjects of works that are just ‘topics’ (e.g., ‘Mural painting and decoration’ which can also be searched under ‘Wall-painting’).
- In FRBROO, F50 Controlled Access Point is declared as a subclass of F13 Identifier, which in turn is a subclass of F12 Nomen. The Latin word ‘nomen’ was used by the FRSAD model to cover signs or arrangements of signs that serve to refer to any kind of thing; it will be taken up by version 2.0 of FRBROO, in replacement of the label ‘Name’. F12 Nomen is declared as a subclass of the CIDOC CRM class E41 Appellation. An instance of F12 Nomen is mentioned in an instance of F35 Nomen Use Statement that forms part of a given instance of F34 KOS (Knowledge Organization System). Normally, instances of F35 Nomen Use Statement should not be arbitrary, but should be based on actual references to instances of F52 Name Use Activity which prove that the nomen is likely to be understood by at least one group of people (e.g., the existence of a book in Russian in which Leonardo da Vinci is named ‘Леонардо да Винчи’ in Cyrillic script bears evidence of the existence of the nomen ‘Леонардо да Винчи’).
- Authority data is to play a crucial part in supporting interoperability between databases of various institutions. It is important to ensure that access points created by museums and libraries for the same notion but which happen to be unidentical are interconnected through references to common URIs. Otherwise, even though we now have a common ontology for libraries and museums, it will be impossible to interrelate, say, the description of a painting in a museum database and the description of a study of that same painting in a bibliographic database. The Linked Open Data Recommendation for Museums available from the ICOM Web site paves the way to the assignment of URIs for museum objects and should be taken into account in any initiative that aims at cooperation of some sort between libraries and museums.
- The notion of event was not absent altogether from the FRBR model, but it was limited, such as the FRBR Final Report was originally written, to events referred to in subject headings.
- In CIDOC CRM, E5 Event is a central notion, which serves to interrelate products, actors, places, and time-spans. The harmonisation process consisted therefore also of introducing events every time they were needed, which contributed to add much precision to the model.
- Since the FRBR model can be said to be work-centred, the first event mentioned in FRBROO is F27 Work Conception, i.e., the event through which a work comes into being. But this is a controversial notion. When does a work begin to exist? In a way, it depends on how we understand the work notion itself. It was decided, in FRBROO, to define the notion of ‘work conception’ as covering the process through which the initial idea for a work occurs in someone’s mind. Not necessarily the mind of the person who does actually create the earliest known expression, however: someone can order a specific work from someone else, who may never have thought of creating such a work before. F27 Work Conception is declared as a subclass of the CIDOC CRM class E65 Creation. The second event encountered in FRBROO is F28 Expression Creation. It is a very peculiar activity, in that it results simultaneously in the creation of a new expression and the production of a physical carrier for that new expression. As a consequence, F28 Expression Creation is declared as a subclass of both E12 Production and E65 Creation. This reflects the fact that F28 Expression Creation consists of two simultaneous and totally inseparable processes, one that affects the intellectual world, and one that affects the physical world. One might think that the next event in FRBROO is something like ‘manifestation creation.’ But it is not so.
- As mentioned earlier, the Manifestation entity was split into two distinct classes in FRBROO: F3 Manifestation Product Type, and F4 Manifestation Singleton. An instance of F4 Manifestation Singleton can be either the carrier of a new, original expression (e.g., an author’s manuscript), or a copy of a pre-existing expression (e.g., a manuscript produced in a mediaeval scriptorium). In the first case, the activity that provokes the existence of that instance of F4 Manifestation Singleton is an instance of F28 Expression Creation; in the second case, it is simply an instance of E12 Production. The creation of an instance of F3 Manifestation Product Type could be modelled as an instance of E83 Type Creation, as F3 is a subclass of E55 Type. But FRBROO also declares a specific class for activities that result in a new publication coming into being. That class is F30 Publication Event.
- The production of all the individual copies of a publication, i.e., of instances of F5 Item, is modelled as an instance of F32 Carrier Production Event, a class that is declared in FRBROO as a subclass of the CIDOC CRM class E12 Production.
- Other types of events were introduced in FRBROO, and I will review them here only briefly.
- F29 Recording Event, a subclass of F28 Expression Creation, corresponds to all kinds of activities that result in recordings (a notion that covers photographs, videotapes and other types of animated images, sound recordings, etc.). F33 Reproduction Event, a subclass of E12 Production, corresponds to all kinds of activities that result in reproductions (i.e., essentially, photocopies, microforms, and digitisations). The difference between a ‘recording’ and a ‘reproduction’ is, in many cases, merely functional and somewhat arbitrary: what differentiates a photograph, on the one hand, from a photocopy, a microform or a digitisation, on the other hand, is that no particular creativity is supposed to be involved in a reproduction process; the resulting object is assumed to be ‘identical,’ as far as certain functionalities are concerned, to the original object.
- F31 Performance, a subclass of E7 Activity, corresponds to all kinds of actions that consist of behaving in a particular way in order to communicate either directly or indirectly to an audience.
- Some events were introduced more specifically in order to address issues relating to authority control.
- F40 Identifier Assignment is basically the same class in FRBROO as E15 Identifier Assignment in the current version of CIDOC CRM (as already mentioned earlier, this is no coincidence, but the consequence of an intentional change in CIDOC CRM). It is an activity that consists of combining signs of any nature (i.e., instances of E90 Symbolic Object) in order to create identifiers (i.e., instances of F13 Identifier or of its subclass F50 Controlled Access Point) that serve to refer uniquely to any instances of E1 CRM Entity.
- When such identifiers incorporate appellations taken from a natural language, their relevance is warranted in authority data by the fact that those appellations are or were actually used by a given group of people, a notion covered by F52 Name Use Activity.
- Two further events correspond to the process through which titles are assigned to instances of F15 Complex Work: the title of a complex work is known through the title of expressions that are deemed to be representative for the work, for which in turn some manifestations are deemed to be representative. For instance, I can know that the title of the work entitled Seitsemän veljestä is Seitsemän veljestä because the expressions that I deem representative for that work are entitled Seitsemän veljestä, and I know that those expressions are entitled Seitsemän veljestä because they are called that way on manifestations that I deem representative for them. If I encounter a manifestation entitled Seven Brothers, or a manifestation entitled Seitsemän veljestä Suomesta, I will not regard the expressions contained in either of these manifestations as representative for the work, and I will not use either of these two titles as an element for the preferred controlled access point for the work (although I can use them for variant access points).
- Presumably, version 2.0 of FRBROO will not be the last one. There is still space for further work.
- Two very important types of resources were only partially modelled in the FRBR model: continuing resources, and digital resources. As a consequence, they are only partially modelled in FRBROO as well, since the principle that was scrupulously followed in the harmonisation process was to take into account the FRBR model such as it stood.
- Issues relating to digital resources are dealt with in CRM dig , an extension to the CIDOC CRM specifically devoted to the provenance of digital objects. I have not investigated in detail if CRM dig addresses all the issues that the original FRBR model left untackled, but it certainly forms a significant starting point. ‘ Continuing resources’ is a generic term for two types of resources that have always been regarded as particularly tricky: serials (i.e., periodicals and bibliographic series), and integrating resources (i.e., both loose-leaf publications and updating digital resources). Although the original FRBR model occasionally mentions some peculiarities of serials, it cannot be said that serials are fully taken into account, which leads to serious difficulties whenever one tries to analyse information about serials strictly within the FRBR framework. Certainly, it would not be very complicated to add some specific classes and properties to FRBROO in order to model title changes, splits, mergers, and other events that make the life of serials so enthralling. When it comes to integrating resources, however, it seems that everything remains to be done. As far as I know, and unless I am mistaken, integrating resources are not specifically accounted for in the original FRBR model, nor in FRBROO. Perhaps they do not need to; perhaps FRBR and FRBROO already tacitly contain all that is required in order to adequately model integrating resources. But it might be worth checking.
- The structural importance of the Work entity is a striking feature of the FRBR model. That fairly abstract notion is linked to the ‘collocating function’ that librarians have always thought their catalogues should perform. The ‘collocating function’ means that users should be enabled to retrieve easily all various editions of all various versions of a given product of the mind, no matter how diverse the titles of those editions and versions can be. Such a preoccupation does not seem to be primary in the museum world.
- Each individual painting or sculpture is regarded in museums as an autonomous piece of art, which may have some privileged relations to other autonomous pieces of art. Even when Giorgio de Chirico reuses the same iconographic elements over and over again, the resulting images are not analysed by art historians and museum curators simply as distinct ‘expressions’ of one ‘work.’ This may explain, I think, why the Work notion was originally absent from the CIDOC CRM, and why it proved at first somewhat difficult to accommodate the Work/Expression dichotomy in that model. The F15 Complex Work class is a useful construct that makes it possible to express a sense of unity among distinct products of the mind, while preserving their own status as works.
- Now that FRBROO exists, what can it be used for? The current buzz word is Linked Data, and obviously FRBROO and CIDOC CRM have a part to play in that domain. Libraries and museums do not need to merge their databases: each type of institution can hold on to the way they view and structure their information; but FRBROO and CIDOC CRM are tools for the integration of the knowledge contained in the databases produced by different types of cultural heritage institutions. On the Semantic Web, it is possible to dream of mediation systems that could process complex queries on museum and library databases. But once again, the necessary condition for the efficiency of such systems and the relevance of returned answers is the accuracy and uniqueness of URIs. FRBROO is an achievement in library and museum cooperation; such cooperation will have made one significant step forward when there is an agreement between museums and libraries on common rules for URI assignment. There is still much to be done, and Linked Data offers us an exciting opportunity for increased cooperation between libraries and museums!